LFM2.5-2.6B
LFM2.5-2.6B is part of LFM2.5, a family of hybrid models designed for on-device deployment. It builds on the LFM2 architecture with a 128K context window and agentic post-training.
Params
2.7 B
Context
131,072
Downloads 30d
188 K
Likes
729
Download history
daily snapshots · 76 days
▲ 0 in the last 30 days (0.0%)
188 K188 K
Oct 12Oct 22Nov 1Nov 10
198 K183 K
Aug 27Sep 21Oct 16Nov 10
Can you run it?
Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.
| File | Quant | Size | Est. VRAM | Verdict on RTX 4090 · 24 GB |
|---|---|---|---|---|
| model.safetensors | bf16 | 5.4 GB | 6.8 GB | ✅ Runs comfortably |
| model.safetensors (bf16, full) | bf16 + 128K ctx | 5.4 GB | 12.9 GB | ✅ Runs comfortably |
Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.
Run it
copy-paste, exact tags checked against the Hub$ curl -s https://aimodelscomparison.com/api/v1/models/lfm2-5-2-6b
{
"hf_id": "LiquidAI/LFM2.5-2.6B",
"params_b": 2.70,
"context_length": 131072,
"license": { "id": "other", "commercial": "unknown" },
"downloads_30d": 187515,
"vram_estimates": [
{ "quant": "bf16", "gb": 6.8 }
],
"updated_at": "2026-08-27T01:00:38Z"
}
Use via API
via OpenRouter · USD per 1M tokens · synced 2026-11-10$0.35
input / 1M
$0.50
output / 1M
$0.04
cached input / 1M
$0.85
1M in + 1M out
:free variant
Free in · Free out
- Context window
- 128,000 tokens
- Max output
- 8,192 tokens
- Providers
- 1
- Cheapest provider
- Liquid · $0.35 / $0.50
- Best throughput
- —
- AA Intelligence / Coding / Agentic
- — / 7.7 / —
✓ Tool calling
✓ Reasoning
✓ Structured outputs
✗ Image input
✗ File input
✗ Audio input
✓ Prompt caching
Specifications
- Architecture
- Lfm2ForCausalLM
- Parameters
- 2.7 B
- Tensor type
- BF16
- Context length
- 131,072
- Vocabulary
- 128,000
- Layers / heads
- 30 / 32
- Licence
- other
- First seen on the Hub
- 2026-07-28
- Training datasets
- undisclosed
- Added to our catalog
- 2026-08-27
Family
Base model and the most-downloaded derivatives in the catalog.
Compare with any text-generation model